The Potential Impact of Health IT

7/27/2019 The Potential Impact of Health IT

1/13

Early Glimpses of the Learning Health Care System:The Potential Impact of Health IT

Summary

In addition to collecting and storing patient inormation or use in individual

clinical encounters, electronic health records (EHRs) provide data or new

types o research and analysis to be undertaken in delivery settings. EHRs en

hance research capabilities by providing data that captures patient outcomes,

is proximate to the point o care, and is available in near real-time. With suchdata, research becomes an important tool in the iterative innovation process

reerred to as the learning health care system. Among the types o inquiry

in delivery systems acilitated by EHRs are quality improvement (QI) analysis

health services research (HSR), analysis to identiy opportunities or workow

efciency, training and research involving simulations, collaborative research

with other organizations, public health research, and new types o clinical

and basic scientifc investigation. The ability to analyze EHR data in delivery

systems has begun to blur traditional distinctions between research, especially

HSR, and QI, creating new opportunities or multidisciplinary innovation

in care delivery and the development o new research methodologies. At the

same time, using EHR data or such inquiry requires particular sensitivityto hardware and sotware capabilities, data quality, and dierences between

cultures o research and health care delivery.

Introduction

The Health Inormation Technology or Economic and Clinical Health

(HITECH) Act provisions o the 2009 American Recovery and Rein-

vestment Act (ARRA, P.L. 111-5) aim to make EHRs and the electronic

exchange o medical inormation the norm in American health care. Th

Ofce o the National Coordinator or Health Inormation Technology

(ONC), the ofce within the U.S. Department o Health and Human

Services (HHS) with primary responsibility or implementing HITECH,is ocused on achieving widespread adoption and initial meaningul use

o EHRs by health care providers and on helping to establish standards

that will support the secure exchange o electronic health inormation.

However, the potential o EHRs goes beyond recording medical inorma

tion about particular patients or use in clinical encounters. EHR data

can be aggregated in various ways and or several purposes. In some

cases, the availability o electronic data enhances activities tradition-

ally undertaken in health care delivery settings such as clinical research,

quality assurance and improvement, and public health surveillance. In

The Health IT or Actionable Knowledge project

examines the experiences o six large health

care systems that have used data rom electronic

health records and other inormation technology

to conduct research and analysis related to

health care delivery. This document is one o

fve reporting the results o this AcademyHealth

initiative. Each report draws on examples rom

these early-adopting health systems to explore a

range o issues relevant to the conduct o health

services and other research using electronic

clinical data. The six health system partners in

this eort are Denver Health, Geisinger Health

System, Kaiser Permanente, the New York City

Department o Health and Mental Hygienes

Primary Care Inormation Project, the Palo Alto

Medical Foundation Research Institute, and the

Veterans Health Administration. AcademyHealth

grateully acknowledges the generous support

o the Caliornia HealthCare Foundation in

unding this project, and the U.S. Agency or

Healthcare Research and Quality (AHRQ) or

providing seed unding.


2/13

the case o health services research, EHRs make the analytic skills

that have heretoore been used mainly in academic pursuits both

possible and valuable or the near real-time delivery o health care

and or the management o organizations that provide that care.1

This report discusses how EHR data are changing the ways in

which we defne and conduct research on health care and the

important new opportunities these electronic capabilities createor health care providers. It is based on a series o meetings and

case studies o six early health IT-adopting health systems in the

United States conducted by AcademyHealth between 2009 and

2011 as part o its Health IT or Actionable Knowledge project.

Three o these organizationsKaiser Permanente, Geisinger,

and the Veterans Health Administration (VHA)are emblem-

atic o the integrated health care delivery systems that were the

earliest adopters o health inormation technology (health IT).

Another two o the organizations studied, Denver Health and

the New York City Department o Public Healths Primary Care

Inormation Project (PCIP), are public health and saety netproviders. The sixth organization, the Palo Alto Medical Founda-

tion (PAMF) and its Research Institute, began as a large multi-

specialty medical group that has merged with other organizations

in northern Caliornia to become an integrated regional delivery

system. Their collective experiences oer insights into how the

health care system as a whole might ully leverage EHR data as

health IT is more widely adopted and used.

How Does Health Information Technology Change

the Potential for Research on Health Care?

Electronic health records, introduced in the 1960s, actually hadtheir roots in research. With a grant rom the U.S. Public Health

Service, Kaiser Permanente tested the frst computer-based medi-

cal record, which it designed to support both patient care and

health services research.2 The VHAs frst EHR system was also

designed by researchers within the organization, who piloted and

studied a prototype EHR during the early 1980s. One reason or

the historical link between research and EHR development may

be the undamental ways in which electronic data expands the

capabilities o researchers while it simultaneously changes the

way patient care is documented. In particular, EHRs acilitate the

availability o data in three important ways:

The availability o clinical data including outcomes data. Tradi-

tionally, electronic data or health services research was largely

limited to administrative claims (health care records submitted

to insurers and other payers by providers or reimbursement

purposes), primary data collected specifcally or research pur-

poses, and vital statistics and data on reportable diseases collect-

ed or public health purposes. Claims data include procedures

and diagnoses. While the coding used in claims is intended to

reect the actual clinical situation or a patient, it is also a tool

in providers strategies to maximize reimbursement, potentially

at the expense o clinical precision and detail. Vital statistics and

surveillance data used by public health ofcials, another source

o data or researchers, also lack clinical detail. Although clinica

laboratories increasingly report results electronically, providers

traditionally store this inormation as part o patients paperrecords. More detailed clinical data collected directly rom

patients retrospectively or research is oten expensive and may

suer rom patients inaccurate memories. Abstracting clini-

cal inormation rom paper medical records is expensive and

requires appropriate training.

By contrast, EHRs provide potentially ready access to detailed

clinical data.3 Although work remains to be done to establish

the relative accuracy o EHR data or particular purposes, EHRs

provide researchers with greater exibility in obtaining clinical

data or a relatively small incremental cost compared to othersources o such data.4

The availability o data proximate to the point o care. Inormation

collected and stored electronically can be aggregated, analyzed,

and provided back to the point o care with ease. Having inte-

grated EHRs proximate to care is not only helpul to the care o

individual patients, but it is also useul to support provider deci-

sions, ensure quality o care, compare provider perormance, and

manage resources at or near the point o care. Traditionally, pro-

viders have had limited inormation available in the clinical care

settingthey have had paper records or individual patients, andthose records did not necessarily contain or provide ready access

to documentation or all test results or care provided to a patient.

Paper records also do not allow providers, at the point o care,

to compare across patients or understand how their care might

dier rom their colleagues. This inormation has the potential to

improve patient outcomes.

The availability o data in near real-time. Creation and prepa-

ration o electronic data or research traditionally takes time.

Administrative data belongs to payers. For Medicare, there

is a lag o two or more years beore claims data are availableto researchers. Similar delays can exist or data rom private

insurers, i they choose to make claims available to researchers

at all. Clinical abstraction o paper records and retrospective

collection o data rom patients also take time. Furthermore,

academic incentives that reward accuracy and certainty over

speed add to the technological barriers that hinder the quick

availability o data and research results.5


3/13

Health IT as a Tool for Rapid Learning and

Innovation

For health care executives and clinical leaders, the most impor-

tant opportunity presented by EHR data may be its central role

in what some experts are calling a learning health care system.6

This strategy uses electronic data to drive a process o discovery

as a natural outgrowth o patient care, ensuring innovation, qual-

ity, saety, and value and serving to reduce the gap between clini-cal care and research.7 In an iterative, rapid-learning cycle, health

care organizations systematically collect and analyze data, use

evidence to identiy opportunities to improve care, implement in-

novations, evaluate the outcomes, and develop new hypotheses to

test. The potential improvements to health care in a learning sys-

tem are seen as coming rom multiple domains, including quality

measurement and improvement, clinical research, and analysis o

the comparative eectiveness o alternative treatments.8

A 2009 Institute o Medicine (IOM) workshop exploring the

application o rapid-learning principles to cancer care ocusedon the variety o inormation tools that will comprise a national

inrastructure or innovation. These tools include interoperability

among data systems or health inormation exchange, patient reg-

istries, databases like the Food and Drug Administrations Sentinel

system (which collects data about adverse events associated with

FDA-approved products), Web-based consumer inormation like

the National Library o Medicines MedlinePlus, and the National

Cancer Institutes open-source Cancer Biomedical Inormatics

Grid (caBIG) described later in the box on page 4. At the base

o this inrastructure or rapid-learning, however, are the EHRs

maintained by individual patients health care providers.

A related ramework or understanding technological change9

highlights many o the types o research undertaken in health care

delivery organizations that are discussed later in this report. In

this ramework, innovation in health care is seen as existing on a

spectrum that ranges rom basic investigation to applied tech-

nological development and diusion. Improvements in health

care delivery begin with basic biomedical research fndings which

are translated into an understanding o the clinical saety and

efcacy o potential treatments or other medical technologies in

a controlled environment through animal studies and humantrials. This knowledge is, in turn, translated into a more thorough

understanding o what types o patients are likely to beneft, and

in what type o setting, through comparative eectiveness and

health services research under real world conditions. Finally, in

order to improve population health, this knowledge is scaled up

and implemented more broadly across the health care system with

ongoing quality measurement and improvement.10

The emergence o clinical research inormatics as a specialized

sub-discipline o the general feld o biomedical inormatics

underscores the centrality o health IT to each o these types

o knowledge translation and the research activities that make

them possible.11 Because EHRs are a major source o research

inormatics, health care delivery organizations are becoming an

institutional home to the ull spectrum o research translation

activities. The organization o research activities at Geisinger

Health Systems among its three research centers, as illustrated

in Figure 1, reects this translational model o research within a

health care delivery organization.

Research in Health Care Delivery Organizations

The six health systems profled as part o AcademyHealths Health

IT or Actionable Knowledge project engage in activities that illus-

trate the range o research and analytic capabilities acilitated by

health IT. Covering the ull spectrum o innovation activities that

comprise a learning health care system as described in the previ-

ous section, the experiences o these organizations also demon-

strate how health IT helps break down traditional defnitions and

boundaries between dierent types o research and analysis. This

section discusses each o these types o research activities, begin-ning with those that can most directly and readily improve the

value o health care services delivered and moving toward those

whose potential to improve care lies mainly in the uture.

Quality Improvement. Health care delivery organizations have long

devoted resources to assuring and improving the quality o care

they deliver. The goal o QI is to eliminate the overuse, underuse,

and misuse o health care services. In its 2001 report, Crossing the

GEISINGER RESEARCH

Basic

Laboratory

Research

Pre-Clinical

Research

Clinical

Trials

HealthOutcomes

Research

Moving

Knowledge to

Practice

Translational Process

Continuity across research spectrum Synergy with clinical enterprise

Patient population and EHR

Center for Clinical

Studies

Weis Center

for Research

Geisinger Center

for Health Research

Figure 1: Translational R&D and the Organization of

Research at Geisigner

Source: Geisinger Health System


4/13

Quality Chasm, the IOM identifed the health care industrys lag

in taking ull advantage o inormation technology as a barrier to

improved quality.12 The access to outcomes data (in addition to the

process inormation that has hereto dominated QI) and its real-

time or near real-time availability, both o which are acilitated by

EHRs, greatly enhance health care organizations ability to address

quality issues. All six o the health systems studied in the Health IT

or Actionable Knowledge project report developing systems to col-lect and assess quality measures and provide immediate eedback to

providers at the point o care through electronic dashboards and

similar tools. Researchers can also use QI measures to study what

aects the outcomes that clinicians believe are important, rather

than the researchers creating their own measures over and over

again. This both speeds the research process and makes the fnd-

ings more relevant.

Health Services Research. According to the Agency or Healthcare

Research and Quality (AHRQ), HSR examines how people get

access to health care, how much care costs, and what happens topatients as a result o this care. HSR seeks to identiy the most e-

ective ways to organize, manage, fnance, and deliver high quality

care; reduce medical errors; and improve patient saety.13 Five o

the six health systems examined or the Health IT or Actionable

Knowledge project provide examples o the diverse ways delivery

systems are integrating HSR into their organizational structures

and missions.14 In each case, the growth in HSR activities is

linked to the capabilities or electronic data collection, storage,

and analysis made possible by EHRs. In summary:

Denver Healths HSR department started as a unit within theCEOs ofce. Originally conceived as a way to identiy and

disseminate best practices and other lessons learned about the

organization and delivery o care in a municipal saety-net

institution to the larger research and delivery system world, this

group has been an active participant in the Accelerating Change

and Transormation in Organizations and Networks (ACTION)

project, a contracted partnership between AHRQ and 15 alli-

ances o delivery systems with robust electronic data capabili-

ties, broad clinical and research experience, and a proven ability

to move research fndings into practice.15 The HSR unit has

subsequently been moved to the Department o Patient Saetyand Quality, reporting to the Chie Quality Ofcer. Topics or

research and analysis are both internally and externally gener-

ated and unded. Members o the HSR department also provide

internal consultation to other Denver Health sta on issues re-

lated to research methods. Denver Healths EHR uses Seimens

sotware.

The U.S. Department o Veterans Aairs Veterans Health Admin-

istration (VHA) maintains an internally unded Health Services

Research and Development (HSR&D) service to undertake research

on patient care, care delivery, health outcomes, cost, and quality in

the VHA. In addition, the service supports the training o clinician

and non-clinician researchers through post-doctoral career develop-

ment awards. HSR&D research occurs throughout the VHAs medi-

cal centers, with locations specializing in particular types or topics

o research. HSR&D developed their frst EHR as a mechanism or

collecting and storing patient data or research. As it has evolved as a

key tool in providing care, the VHA has also enhanced the ability to

National Cancer Institutes Cancer Biomedical

Infomatics Grid (caBIG)Between 2004 and 2010, the National Cancer Institute

(NCI) has invested more than $350 million in the Cancer

Biomedical Informatics Grid (caBIG), a collaborative IT

infrastructure for data collection, integration, analysis, and

dissemination across NCI centers and programs designedto facilitate the discovery of new approaches to detection,

diagnosis, treatment, and prevention of cancer. Begun as an

attempt to develop standards for interoperability and analytic

tools for cancer researchers and as a forum for comparing

data related to gene expression related to cancer research,

caBIG was expanded in 2007 to an initiative to develop a

comprehensive open-source enterprise system to support

all aspects of cancer research. Plans for this system have

included an EHR and cloud computing.16

Because of caBIGs ambitious scope and signicant budget,NCI Director Harold Varmus, M.D., created a working

group to review the program upon his appointment in

July 2010. The groups report underscored some of the

potential problems in developing broad health IT systems

from scratch. The working group afrmed the relevance

of caBIGs original goals. In particular, caBIG moved the

cancer research community beyond messaging systems

and limited structured vocabularies to an infrastructure that

allows data to be harmonized across cancer centers.

However, it strongly criticized the programs effort to expandbeyond those goals to develop an overly complex and

ambitious software enterprise of NCI-branded tools,

especially for managing clinical trials. The report concluded

that these NCI tools duplicate established commercial

software, have not been widely adopted, and do not

provide benet commensurate with the upfront and on-

going investment they require. The working group saw

the lack of independent oversight and non-peer-reviewed

funding decisions as key to caBIGs difculties. In addition to

recommending that NCI correct these short-comings in its

process, the working group suggested that caBIG return toits original goals.17


5/13

use EHR data or quality improvement by developing regional and

national data warehouses and creating VA Inormatics and Comput-

ing Inrastructure (VINCI), a secure virtual environment to improve

researchers use o VHA data while ensuring patient privacy and

data security. The VHA uses EHR sotware it created itsel.

Kaiser Permanente (KP) is actually three distinct organizations:

Kaiser Foundation Health Plan, Kaiser Foundation Hospitals,and the Permanente Medical Groups. Together, they operate

eight regional health care organizations rom Hawaii to Wash-

ington, D.C. The two Caliornia regions (North and South)

account or two-thirds o the KP membership and nearly all

the Kaiser Foundation Hospitals. KP was a pioneer in capi-

tated health coverage in the United States, and its core product

oerings are still ull-service HMO plans. In response to the

evolving health care marketplace, however, KP now also oers

a variety o high-deductible and sel-unded plans to purchas-

ers. Most o the KP regions have dedicated research units that

undertake both internally and externally unded studies. Theseresearch units include the Division o Research in the Northern

Caliornia region, the Department o Research and Evaluation

in Southern Caliornia, the Institute or Health Research in

Colorado, and the Center or Health Research which includes

the Georgia, Hawaii, Northwest and Mid-Atlantic regions. In

addition, the KP Center or Eectiveness and Saety Research

(CESR), ounded in 2009, represents a national network o re-

search across all eight Kaiser regions. HSR, which includes both

institutional- and investigator-initiated studies, represents a

signifcant portion o the more than 1,000 researchers and sta

and the $140 million annually (in 2010)18

devoted to research byKP. One distinctive eature o KP as a venue or HSR is that it

unctions as a capitated insurer and payer as well as a provider

o care, which gives the organization a particular incentive to

achieve greater value or each dollar in care delivered. KPs EHR

uses Epic sotware.

Geisinger is an integrated health care delivery system serv-

ing 31 o Pennsylvanias 67 counties, mainly in the central and

northeastern parts o the state. The system sees about 350,000

primary care patients and about 700,000 specialty care patients

each year. Scientifc investigation at Geisinger is built upona translational research and development model in which

research is part o a continuum running rom basic research

to clinical trials and outcomes research to the implementation

o new knowledge into clinical practice. The Center or Health

Research, started in 2003, houses HSR as well as epidemiologic

and community health research. The Clinical Innovations team,

oten in collaboration with the Center or Health Research,

implements new models o care, and Geisinger Ventures seeks

commercial partnerships to develop and market Geisinger in-

novations or the larger health care system. Spending or HSR is

not broken out separately rom the $16 million in total research

spending at Geisinger, about 55 percent o which is supported

by external or endowment unds (i.e., not clinical practice or

other reimbursed care). Geisinger has its own capitated health

plan covering 230,000 individuals, about hal o whom get mosto their care rom Geisinger. Like KP, the role o insurers may

increase the value o HSR and general research since the health

system has a larger incentive than do providers without insur-

ance risk to seek greater value through improved quality and

efciency. Geisingers EHR uses Epic sotware.

The Palo Alto Medical Foundation Research Institute (PAM-

FRI), which was ounded in 1950, has long conducted HSR

including some o the earliest studies on the cost o care in the

1960s.20 PAMFRI is the dedicated research unit o the Palo Alto

The HMO Research Network (HMORN)

Three of the health systems examined as part of

AcademyHealths Health IT for Actionable Knowledge

project (Geisinger, KP, and PAMF) are members of

the HMORN. In addition to serving as a forum for

researchers at member organizations to share ideas and

best practices, the HMORN provides the infrastructureto carry out collaborative studies in epidemiology,

comparative effectiveness, and other health services

research. Support for HMORN research can come from

the health plans themselves or from external sources,

including the NIH Collaboratory established by the NIH

Common Fund to facilitate the translation of research

ndings into patient care.

Key to the HMORNs research infrastructure is its ability

to draw on the EHRs of its member health plans. The

HMORN has created a virtual data warehouse (VDW)consisting of patient level administrative and EHR

data. Using a set of standards established by an

HMORN-wide working group, member health plans

have created a parallel set of databases of pre-dened

variables. Creating the databases ahead of time helps

assure the efciency of the process and the quality

of the data. Maintaining the data at each health plan

minimizes threats to data security and privacy. The

VDW comprises demographic, health plan enrollment,

encounter, procedure, diagnostic, provider, cancer/

tumor, pharmacy, vital sign, and laboratory data.Because multi-center research adds to the regulatory

complexity of obtaining approval to use data, the

HMORN has also established streamlined procedures

for creating data use agreements and for IRB review.19


6/13

Medical Foundation (PAMF), a multi-specialty group medical

practice o more than 900 physicians and about 600,000 patients

in our northern Caliornia counties. PAMF introduced an EHR

using Epic sotware in 2000, and since 2008, all o PAMF uses

this one system. PAMFRI reinorced its commitment to HSR

by recruiting a senior health services researcher rom academia

to become the organizations director in 2008. Through close

working relationships with clinical and executive leaders oPAMF, health services researchers at PAMFRI identiy and ad-

dress practical research questions with the intent o improving

quality and efciency o care. With appropriate privacy and pa-

tient protections, researchers have access to a regularly updated

copy o the EHR database. Though some work is internally

supported, researchers seek external unding as appropriate. Re-

gardless o the source o support, the researchers give priority to

questions that can result in publishable, generalizable knowl-

edge. O the $9.6 million budgeted or HSR in 2012, 69 percent

is rom external sources with the remainder rom other income,

gits, return on investments, and PAMF itsel.21

A separate report written under the auspices o this project,

HSR Agenda Setting: Lessons rom Three HIT-Enabled Health

Systems, examines the HSR unction at Denver Health, Geisinger,

and PAMFRI in greater detail. In particular, it examines the his-

tory, placement, and role o HSR in each health system, how each

organization determines what HSR questions to pursue, and the

sources o HSR unding.23

Research and Analysis or Efciency Improvements. EHRs and

related health IT can also help support delivery organizations

eorts to increase value by making the process o care more

efcient. In recent years, Denver Health has adopted the Lean

methodology, an approach originally developed by the automo-

tive manuacturer Toyota, to reduce waste and improve the health

care experience or patients. In an analysis o their Lean experi-

ence undertaken or AHRQ, Denver Health cites the beneft o

inormation technology to provide the data needed or baseline

and on-going monitoring.24 The box to the let discusses the use

o Lean at Denver Health in greater detail. In the course o theHealth IT or Actionable Knowledge project, Geisinger, KP, PAMF,

and the VHA all also cited the importance o clinical and admin-

istrative electronic data to support eorts to reduce waste.

Simulation in Research and Training. Simulation in health care has

emerged as a signifcant tool or minimizing medical error, im-

proving health outcomes, and creating efciency. For some health

care services in which experimenting on or learning rom real pa-

tients puts those patients well-being at risk, simulation can oer

an eective alternative. Simulations can be used or training or to

explore alternative clinical or management decisions. They cantake several orms, varying in complexity, approximation o real-

ity, and technological ormat. They can present the patient cases

or other situations in verbal ormat, using actors or dummies in a

realistic setting, or with computers. Outcomes can be determined

by established rules and probabilities or by expert evaluation.25

Data rom EHRs can acilitate simulations by providing the

knowledge base to identiy areas where simulations may improve

provider perormance and provide the basis or understanding

the likely outcomes o simulated actions. In the past several years,

research involving simulation has become a unding priority or

AHRQ.26 Several o the health systems examined or this project,including Geisinger, KP, and the VHA, have developed simulation

capabilities or training, evaluation o technology, and research.27

Collaborative Research. Another trend among delivery systems

with EHRs is their increasing involvement in research that spans

multiple organizations. Collaboration in clinical, health services,

or other research drawing on patient experiences among dierent

health care organizations can increase sample sizes and provide

opportunities to examine a greater diversity o patient popula-

tions. One key to such collaboration is the ability to understand,

The Use of Lean Process at Denver Health

In 2005, with initial support from AHRQ, Denver

Health introduced the Lean method for rapid-cycle

improvements to eliminate waste from the process

of delivering health care. Based in part on the quality

improvement theories of statistician W. Edwards Deming,

Toyota rst developed Lean for application to automobile

manufacturing. The Lean process attempts to distinguish

those steps in an organizations work ows, or value

streams, that add value for patients from those that do

not. In adapting Lean to health care delivery, Denver

Health relies on Rapid Innovation Events (RIEs), in which

staff examine a particular value stream, nd opportunities

for greater efciency with the goal of eliminating 50

percent of the waste, and implement appropriate

changes, all within a one-week period. Managers

and clinical staff participate in several RIEs each year.

Administrators and clinicians throughout the organization

who receive special training to become Lean Black Belts

are responsible for identifying additional opportunities to

eliminate waste. A key component of the Lean process

is the identication of metrics and data with which to

evaluate the impact of changes to a given workow.

Between 2005 and 2009, the Lean process generated

$42 million in nancial benet to Denver Health with $8.8

of that amount attributable to the Black Belts alone. The

program has gained momentum over time with over half

of the $42 million realized in 2009 alone.22


7/13

harmonize, and possibly exchange data among dierent organiza-

tions EHRs. Another key is the ability or individual researchers

to bridge cultural or other dierences across organizations and

among themselves. As discussed later in this report, these require-

ments can present signifcant challenges.

All regions o KP share common EHR sotware, although data

are not routinely shared across those regions. Eorts o the KPCenter or Health Research, which undertakes studies that can

span the health plans northwest, southeast, and Hawaii regions,

represent one such eort. Similarly, each VHA medical center

uses the same EHR but maintains its own database o patient re-

cords. However, the creation o regional and national data ware-

houses has acilitated research involving more than one medical

center. The Health Maintenance Organization Research Network

(HMORN), a consortium o 19 health care delivery organizations

including KP, Geisinger, and PAMF, provides a more extensive ex-

ample. At the core o HMORN is a virtual data warehouse com-

prising a set o standardized data ormats and coding conventionsused by HMORN members. These conventions allow each plan

to extract standardized data fles or use by researchers in particu-

lar studies. The box on page 5 discusses the role o electronic data

in HMORN activities in greater detail.

Research to Support Public Health Functions. EHRs also provide

a new tool or those charged with public and population health.

By electronically querying the EHR systems o individual provid-

ers, the New York City PCIP has been able to conduct near real-

time snydromic surveillance, in which the city is able to measure

the number o patients presenting at their physicians ofces withparticular symptoms (e.g., u-like symptoms) to be able to track

the potential spread o inectious disease or environmentally-trig-

gered health problems by neighborhood. In addition, PCIP can

track progress toward achieving city-wide goals or prevention

such as or immunizations, disease screening, or chronic disease

management. Such inormation provides a potential tool or

identiying and addressing public health needs more quickly than

do more traditional, labor-intensive reporting and primary data

collection. It also suggests that EHRs provide an opportunity or

primary care providers to pursue public health objectives when

treating individual patients. A separate Health IT or ActionableKnowledge report, Using Health Inormation Technology to Im-

prove Health and Health Care in Underserved Communities: The

Primary Care Inormation Project, examines New York Citys

experiences in greater detail.28

Clinical Research. Clinical research, especially research testing the

saety and eectiveness o new pharmaceuticals, other therapies,

and diagnostics, has long had its own inrastructure including

tools or the collection, storage, and analysis o research data.

Five o the health systems partnering on this project participate

in clinical trials. The New York City PCIP, part o a public health

department, does not.

Much clinical research, especially studies sponsored by pharma-

ceutical or medical device companies, have been managed by

frms known as clinical trial organizations (CTOs) or contract

research organizations (CROs) that have traditionally collectedand analyzed research data with their own sotware and computer

systems. The adoption o EHRs by hospitals and other delivery

organizations participating in clinical research provides an op-

portunity to integrate data or research and care unctions. Five

o the six partnering health systems or this project carry out

clinical research.29 The extent to which data collection, storage,

and analysis or clinical research is integrated into these systems

varies across and within institutions. In general, clinical research

studies managed by outside CTOs/CROs tend to have their own

electronic data systems while research studies initiated and man-

aged internally are more likely to integrate their data collectionand management to some degree with the EHR, or example by

collecting needed patient data directly rom the EHR or storage

or analysis o research data behind the same electronic frewall

that protects EHR data.30 Among health systems examined as part

o this project, the VHA is currently building capacity to leverage

its EHR in the conduct o clinical trials sponsored by industry or

other outside unders.31

Although such integration provides potential efciencies, it also

presents many o the technological, methodological, and gov-

ernance issues briey described below. A separate report romAcademyHealths Health IT or Actionable Knowledge project,

Finding Value in Volume: An Exploration o Data Access and

Quality Challenges, explores issues o data inrastructure, data

quality, and data governance as experienced by the six health sys-

tems in greater depth. As described in the box on page 4, a recent

initiative o the National Cancer Institute, the Cancer Biomedical

Inormatics Grid (caBIG) illustrates both the potential and chal-

lenges o using integrated electronic systems to support clinical

research.32

Basic Research, Genomics, and Phenomics. Basic scientifc investi-gation to better understand human physiology, genetics, disease,

and the basis or potential new treatments has traditionally been

the purview o universities, academic medical centers, and to the

extent that it provides the basis or potential new diagnostic tools

or treatments, the laboratories o pharmaceutical and biotechnol-

ogy frms. Computer inormatics is a key tool in gene sequencing

and related genomic research, even though the EHRper seis o

limited value to basic science.33


8/13

At the same time, however, EHRs can be particularly useul in

translational research that tries to bridge basic biological under-

standing and human health. Phenomics, the study o an organ-

isms physical and biochemical characteristics and how changes

in genetic make-up and environmental actors aect them, is an

example o this type o scientifc investigation. Such research is

ultimately intended to be the basis or personalized medicine,

defned as the tailoring o medical treatment to the individualcharacteristics o each patient.34 In the delivery o care, a link to

a patients genetic inormation known to be associated with par-

ticular health conditions may help clinicians provide appropriate

care. Such research links data rom biological samples, especially

genetic material, with clinical data recorded in patients EHRs.

Several o AcademyHealths partners on the Health IT or Action-

able Knowledge project are leveraging their EHRs or phenomic

and genomic investigation, including a large project by KPs

Northern Caliornia region in collaboration with the University o

Caliornia, San Francisco. With $25 million in support rom theNational Institutes o Health, KP is creating a large data repository

o more than 400,000 health plan members to support studies o

genetic associations with drug metabolism and response, disease

progression, development, and recurrence, environmental inorma-

tion, as well as characteristics o patients liestyle and behavior. 35

The VHA is also creating a large genomics cohort called the Million

Veterans Program, which makes use o that health systems EHR.36

New Opportunities and Challenges

The availability o electronic data gives health systems new ways

to use research to better understand their organizations and thecare they provide, but at the same time, these new capabilities

have signifcant implications or the research process itsel.

The Blending o HSR and QI. As more health services research-

ers take advantage o EHR data to work directly with delivery

systems, the line that has traditionally distinguished HSR rom

other types o measurement and analysis that support health care

administration and care has become less relevant. In the case o

QI, or the purposes o privacy and human subject protections,

the traditional distinction is that QI is part o the management

o health care delivery systems, while research is perormed inorder to produce generalizable knowledge.37 However, or health

services researchers who work in delivery organizations, their

research agendas are oten intertwined with the work o their QI

colleagues. For example, as described by one delivery system-

based health services researcher providing input to this project,

his HSR agenda includes assessing and describing the generaliz-

able lessons that can be gleaned rom practice changes initiated

by QI proessionals, and raising, rom a research perspective,

questions whose answers could be readily applied by the operat-

ing organization. The QI community is also increasingly ocused

on systematically evaluating processes or assuring and improving

quality as evidenced by the emergence o new felds o inquiry like

improvement science and other attempts to evaluate and learn

rom actual innovations in QI. As discussed earlier, both health

services researchers and QI proessionals make use o EHR data

that is available in near real-time close to the point o care.

The health systems examined as part o the Health IT or Action-able Knowledge project noted several benefts and implications o

this blurring o QI and research:

Regular interactions among health services researchers, experts

in operations research, QI proessionals, clinicians, IT special-

ists, and other proessionals at these institutions yield multidis-

ciplinary interpretations o problems and data, and are key to

developing innovative approaches to achieving better cost and

quality outcomes.

The presence o researchers on the ront lines o care delivery has in-troduced approaches to methodology that can run counter to tradi-

tional academic norms. As described by one delivery system-based

health services researcher, academic rewards are oten weighted

towards sophisticated or new methods and dramatic answers to

what are oten narrowly defned questions. In delivery systems,

however, greater value is given to more widely applicable results

produced more quickly. Health services researchers involved in

this project also noted that there can be dierences in the necessary

level o certainty or academic and health services research. They

suggested that this dierence may be related to the degree o control

maintained over the results. In traditional academic research, theresearcher has little control over how results are used once they are

openly published. In addition, academia generally rewards the use

o sophisticated methodological approaches. As a result, the peer-

review process puts substantial emphasis on achieving a high level

o certainty and identiying limitations. By contrast, delivery system

researchers retain signifcant control over how their results are used.

I analysis suggests a particular course o action, the organization

can implement it. I a particular innovation does not work, the

organization can abandon it or try an alternative. Everything else

being equal, having such control over how research is used may

reduce the level o certainty necessary to act.

Even i the standards o evidence or delivery system research can

vary rom those expected in academia, researchers who work on

the ront lines o health care have noted the need or new analytic

methods and inquiry appropriate to the use o EHR data.38 Meth-

odological challenges include fnding new ways to deal with biases

and conounding variables common to research not based on ran-

domized controlled trials (RCTs) and developing better approaches

to replicating results rom one setting to another or scaling interven-

tions up rom a pilot phase to ull implementation.


9/13

Data Inrastructure. A key potential beneft o delivery system

research involving EHR data is the ability to link data across health

care organizations. The six health care systems that participated

in the Health IT or Actionable Knowledge project cited several

reasons why such interconnectedness is desirable. Examining

data rom more than one organization or site o care can increase

sample sizes, making it easier to detect hypothesized eects. In

addition, multiple locations can create opportunities to examinenatural experiments or isolate geographic or institutional actors

related to outcomes o interest. One health system cited the ability

to link to data beyond their own organization as an important tool

in recruiting the best health services researchers. Another health

system saw the ability to link to other organizations data as an op-

portunity to learn about the benefts and drawbacks o the hard-

ware and sotware systems that other health systems have chosen to

inorm their own organizations uture purchasing decisions.

In order to use data rom dierent health care delivery organiza-

tions, however, there has to be an inrastructure that allows dierentcomputing systems to interact and exchange data. Even when two

organizations use the same sotware, it is possible that they record data

in dierent felds or defne particular variables dierently. Such varia-

tion can occur even within a single organization. In order to use data

rom dierent systems, there need to be standards established beore

the research is undertaken, or researchers need to invest resources to

understand and harmonize the data so observations are comparable.

Deciding what data to collect can also present difculties. Data needs

or research can dier rom those or clinical care or administrative

operations. For example, most clinical care can occur even i neces-sary inormation is in ree-orm text notes or scanned images o

non-digitized records. Data in these ormats, however, represent a

signifcant hurdle or research. In addition, or retrospective study,

the variables needed or HSR or QI may not be in the EHR. Even or

prospective research or analysis, clinicians only have a fnite amount o

time to record data during a patient encounter, presenting potentially

difcult choices about what pieces o inormation are most impor-

tant to collect.39 Even i the data needed or a particular study is not

readily available, electronic systems generally create large amounts o

data not previously available. The volume reects both the number

o observations (i.e., patients) in a database as well as the number odata elements available or each o those observations. Health systems

participating in this project noted that the potential or large volumes

o data to overwhelm researchers underscores the value o advance

planning and the involvement o researchers in the initial design and

implementation o data inrastructure.

Data Quality. Although EHRs acilitate the use o data or research

and analysis, they oten require more time and resources to clean than

do administrative or other data used by researchers in the past. Health

systems researchers involved in this project pointed out that this need

does not reect less accuracy in EHR data than in traditional research

data; rather, EHR data provides better opportunities or researchers to

analyze its quality and clean it as necessary.

Threats to the quality o EHR data can arise rom several sources.

There can be inconsistencies in how oten or in what felds dier-

ent clinicians enter data.40

As mentioned in the discussion o datainrastructure, there can be variation in how dierent clinicians

interpret particular variables. Another potential difculty derives

rom the use o open text felds in which clinicians can record

notes in ree orm as opposed to using defned felds. Inormati-

cians are developing sotware or Natural Language Processing

(NLP), which attempts to electronically interpret ree-orm text

in order to extract useul inormation. However, such sotware is

still in development and varies in its eectiveness.

As with methodological rigor, accuracy is desirable, but the level

o data quality necessary can depend on its use, particularly whenachieving greater accuracy requires additional time and resources.

For clinical or administrative decision-makers on the ront lines

o health care delivery, the cost o not having timely inormation

may be greater than the beneft o achieving greater confdence in

the accuracy o the data. For traditional research, the incentives

are oten reversed. Working out what level o data accuracy is

needed or what purpose is an on-going process or researchers.41

A separate Health IT or Actionable Knowledge report examines issues

related to data inrastructure and data quality in greater detail.42

Data Governance. As suggested above, the regulatory require-

ments or human subjects and privacy protection are dierent,

and generally more restrictive, or research uses o data than they

are or QI activities, which are considered part o patient care.

The blurring o lines between research and QI has created some

uncertainty about appropriate data governance. While the basic

concepts and rules, including the Common Rule,43 have not

changed, discussions with the six health systems examined in this

project suggest that compliance with those rules may become

more complicated. For example, i a researcher wants to explore

a change in clinical care, this typically requires a ull IRB reviewand inormed consent by the patient/subjects. I a clinic wants to

change its care, it is considered QI and no IRB review or inormed

consent is required. What happens, however, i a researcher wants

to evaluate the clinics decision to change practice? Is IRB review

required? I an IRB does not review the researchers eort, many

journals will reuse to consider the resulting papers or publica-

tion. I the IRB is asked to review the intervention, is a ull con-

sent required o the patients, even i may impede workow and

make the clinic unwilling to make the change?


10/13

Discussions with dierent health systems also suggest that there is

institutional variation in how they interpret and implement human

subjects guidelines. For example, one health system participating in

the Health IT or Actionable Knowledge project requires all investi-

gation to be reviewed by their IRB. However, the organization has

made the IRB process relatively simple or minimally-risky inquiry.

Another health system, citing a lengthy and cumbersome review

process and IRB members generally unamiliar with non-clinicalresearch, reported a more liberal interpretation o what analysis

requires IRB review. Researchers rom all o the health systems

agreed that in light o new research capabilities created by health

IT, the rules, guidance, and processes related to human subjects and

privacy protection need updating, especially as they apply to HSR

and other non-clinical trial research.

A separate Health IT or Actionable Knowledge report, Legal and

Policy Challenges to Secondary Uses o Inormation rom Elec-

tronic Clinical Health Records, looks specifcally at issues related

to data governance.44

Bridging Cultural Divides. Undertaking HSR and other research in

health systems also requires sensitivity to the dierences between

traditional academic culture and that o health care delivery.

The need or aster turn-around and the acceptability o more

uncertainty in the results in delivery system research have already

been mentioned. Another dierence is the degree o collabora-

tion expected in the research process. In traditional academic

HSR, the scientifc culture is more oriented toward the individual

researcher.45 Typically, an investigator identifes an interest-

ing question, fnds appropriate data, secures external unding,and when satisfed with the validity o the results, disseminates

them through peer-review publication and scientifc conerences.

Academic researchers may partner with a delivery system, but

this is usually to obtain access to theirdata and many researchers

eel that as long as they abide by the data use agreement, they have

ulflled their obligation to the delivery system. Although one

project may beget another, it is ar more common that at the end

o the project (or even the completion o the data transer) there

is no urther communication with the delivery system.

In contrast, researchers who choose to workwithindeliverysystems have strong reasons to nurture their relationship to the

organization, even i their work is totally externally unded. Data

by itsel is useul; data with access to the people who created it,

who can explain its nuances, and who can provide additional

inormation is extraordinarily valuable and oers the researcher

a competitive edge in external unding. Providing such data and

the access to the human capital behind it is costly to the delivery

system, but is a cost well worth bearing i the organization can

see some returns rom its collaboration with the researchers. The

implications o these cultural dierences include:

Health services researchers in the organizations examined

as part o the Health IT or Actionable Knowledge project

stressed the importance o researchers developing both per-

sonal relationships with clinicians and administrators and anunderstanding o the incentives and pressures they ace in order

to identiy research projects o value to the organization. They

also stressed that personal relationships with clinicians can be

key to understanding how they record inormation in the EHR

and in interpreting research results.

The graduate programs that train new health services research-

ers could serve the feld by developing new curricula and practi-

cal experiences that amiliarize young health services research-

ers interested in working in or with delivery systems with how

these organizations operate. For mid-career researchers, therewould also be value in developing learning experiences that pro-

vide a hands-on understanding o delivery system operations

and the organizational values that underlie those operations.

A commitment by delivery systems to HSR can also require

adjustment or clinicians and administrators. For providers

and administrators already acing competition or their time

and attention, the potential beneft o research and working

with researchers may not be immediately apparent. Devot-

ing resources to research can be seen as taking resources away

rom patient care. One area where this tension has played outat some o the health systems examined as part o the Health IT

or Actionable Knowledge project has been in access to data and

IT proessionals. The experience o these same health systems,

however, suggests that leadership and a vocal commitment to

research rom the corporate suite can help other proessionals

in the organization appreciate its value.

Conclusion

The experience o the six institutions examined as part o Acad-

emyHealths Health IT or Actionable Knowledge project confrms

the potential value o electronic data systems or multiple usesbeyond patient record keeping. However, by examining only six

health systems, this project can only provide a avor o the benefts,

costs, risks, and challenges associated with secondary, analytic uses

o EHRs. More eort is needed to identiy and disseminate best

practices and to know how to translate them or the great diversity

o delivery organizations that will eventually have the capacity to

use their EHR systems or more than just documenting patient


11/13

care. Nonetheless, the experience o these early health IT-adopting

entities can serve to sensitize both researchers and health care deliv-

ery organizations to the ways in which electronic data changes the

way research and analysis o health care is evolving.

About the Authors

Michael E. Gluck, Ph.D., M.P.P., is the Director o Translation

Strategies at AcademyHealth. Megan Ix, B.S., Associate at Acad-

emyHealth, and Bryan Kelley, B.A., Research Assistant at Acad-

emyHealth, provided research assistance or this report.

Acknowledgements

AcademyHealth grateully acknowledges the time and exper-

tise provided in the preparation o this report by the research-

ers, clinicians, IT proessionals, executives at the health systems

participating in AcademyHealths HIT or Actionable Knowledge

project Denver Health, Geisinger Health System, Kaiser Perma-

nente, the New York City Department o Healths Primary Care

Inormation Project, the Palo Alto Medical Foundation ResearchInstitute, and the Veterans Health Administration. Any errors are

AcademyHealths.

Endnotes1 Mandl, K.D. and T.H. Lee Integrating Medical Inormatics and Health Services

Research: The Need or Dual Training at the Clinical Health Systems and

Policy Levels, Journal o the American Medical Inormatics Association, Vol 9,

No. 2, March/April 2002, pp. 127-132.

2 Kaiser Permanente. History o the Division o Research. Retrieved rom http://

www.dor.kaiser.org/external/DORExternal/about/history.aspx, accessed on

January 27, 2012.

3 Weiner, G. and P.J. Embi. Toward Reuse o Clinical Data or Research

and Quality Improvement: The End o the Beginning? Annals o Internal

Medicine, Vol. 151, No. 5, September 1, 2009, pp. 359-360.4 Researchers may not necessarily want data that perectly captures the true

clinical measurement or a patient. I they are trying to generalize about

actual medical practice, they may preer EHR data recorded by clinicians in

the course o a regular patient encounter, even though that data may include

inaccuracies that reect the noise o actual practice. In other cases, the

researcher may actually need to know the true clinical value or a patient,

which may require a prospective data collection plan that goes beyond usual

practice. The important point here is that EHRs oer the researcher new

options or obtaining clinical data.

5 One exception in clinical trial research occurs when ethical considerations

require that clearly better treatments be made available to all research subjects

as soon as eectiveness and saety are established. However, the results o most

trials are made available ater research subjects are treated.

6 Etheredge, L.M., A Rapid-Learning Health System Health Aairs, Vol. 26, No.

2, January 26, 2007, pp. w107-18; Friedman C.P., et al. Achieving a Nationwide

Learning Health System, Science Translational Medicine, Vol. 2, No. 57,

November 10, 2010, pp. cm29-31.

7 Abernathy, A.P. et al. Rapid-Learning System or Cancer Care, Journal o

Clinical Oncology, Vol. 28, No. 27, September 20, 2010, pp. 4268-74.

8 Etheredge, LM, 2007, op.cit.; Abernathy, AP, et al, 2010, op. cit.

9 Godin, B, The Linear Model o Innovation: The Historical Construction o

an Analytic Framework, Science, Technology & Human Values. Vol. 31, No. 6

(November 2006), pp.639-667.

10 Dougherty, D. and P.H. Conway, The 3Ts Road Map to Transorm U.S.

Health Care: The How o High-Quality Care, Journal o the American

Medical Association, Vol. 299, No. 19, May 21, 2008, pp. 2319-21. The

translation o basic biomedical understanding to clinical efcacy knowledge

is sometimes reerred to as T1, the translation o clinical efcacy under ideal

conditions into an understanding o clinical eectiveness in the real world

as T2, and the translation o clinical eectiveness knowledge into improved

health care quality and population health as T3.

11 Embi, P.J., et al. Clinical Research Inormatics: Challenges, Opportunities

and Defnition or an Emerging Domain, Journal o the Medical Inormatics

Association, Vol. 16, No. 3, May/June 2009, pp. 316-27.

12 Committee on Quality o Health Care in America, Institute o Medicine.

(2001). Crossing the Quality Chasm: A New Health System or the 21st

Century. Washington, DC: National Academy Press, pp. 31-4.

13 Agency or Healthcare Research and Quality, U.S. Department o Health and

Human Services. (2002). Helping the Nation With Health Services Research.

Fact Sheet. Rockville, MD. Retrieved rom http://www.ahrq.gov/news/ocus/

scenarios.htm accessed on January 27, 2012.14 As mentioned earlier, the sixth health system, the NYC PCIP, which is part o

the citys Department o Health and Mental Hygiene, ocuses on public and

population health.

15 AHRQ competitively awards contracts among participating ACTION teams on

topics relevant to the practice, organization, and management o health care

delivery. The ACTION contract lead by Denver Health also includes saety

net institutions rom Baltimore, Minneapolis, New York City, Dallas, and the

University o Colorado Hospital.

16 National Cancer Institute. (2009). The Cancer Biomedical Inormatics Grid

caBIG Resource Guide (NIH Publication No. 10-7518).

17 National Cancer Institute Board o Scientifc Advisers, 2011, op cit.

18 Davis, R.L. (2010). KP Center or Eectiveness and Saety Research. Slides

rom presentation to the Kaiser Permanente Center or Health Research SE,

September 13, 2010. Retrieved rom http://www.c-path.org/pd/DavisPSSW.

pd , accessed on January 27, 2012.

19 HMO Research Network. (2006). Collaboration Toolkit: A Guide to

Multicenter Research in the HMO Research Network. Retrieved rom

http://www.hmoresearchnetwork.org/resources/toolkit/HMORN_

CollaborationToolkit.pd , accessed on January 27, 2012.

20 See or example, Scitovsky, A.S. Changes in the Costs o Treatment o Selected

Illnesses, 1961-65, American Economic Review, Vol. 5, No. 57, December 1967,

pp. 1182-95.

21 PAMFRI Director Hal Lut, personal communication, November 3, 2011.

22 Agency or Healthcare Research and Quality. (2007). Managing and Evaluating

Rapid-Cycle Process Improvements as Vehicles or Hospital System Redesign

(AHRQ Publication No. 07-0074-EF). Rockville, MD: retrieved rom http://

www.ahrq.gov/qual/rapidcycle/, accessed on January 27, 2012 and Agency or

Healthcare Research and Quality. (2002). A Toolkit or Redesign in Health

Care (AHRQ Publication No. 05-0108-EF, Prepared by Denver Health under

Contract No. 290-00-0014). Rockville, MD: retrieved rom http://www.ahrq.

gov/qual/toolkit/, accessed on January 27, 2012.23 Pittman, P. HSR Agenda Setting: Lessons rom Three HIT-Enabled Health

Systems, Health IT or Actionable Knowledge report, AcademyHealth,

February 2012.

24 Agency or Healthcare Research and Quality. (2007). Managing and Evaluating

Rapid-Cycle Process Improvements as Vehicles or Hospital System Redesign

(AHRQ Publication No. 07-0074-EF, September 2007). Rockville, MD:

Retrieved rom http://www.ahrq.gov/qual/rapidcycle, accessed on January 27,

2012.

25 Okuda Y et al. The Utility o Simulation in Medical Education: What is the

Evidence? Mount Sinai Journal o Medicine, Vol. 76, No. 4, August 2009, pp

330-43; Aliner, G. A Typology o Educationally Focused Medical Simulation

Tools Medical Teacher, Vol. 29, No. 8, 2007, pp. e243-8; Bond, W.F., et al. The

Use o Simulation in Emergency Medicine: A Research Agenda, Academic

Emergency Medicine, Vol. 14, No. 4, April 2007, pp. 353-63; Patterson, M.D.

et al. In Situ Simulation: Challenges and Results in Agency or HealthcareResearch and Quality. (2008). Advances in Patient Saety: New Directions and

Alternative Approaches (Vol 3: Perormance and Tools) (Publication No.: 08-

0034-3). Rockville, MD: Retrieved rom http://www.ncbi.nlm.nih.gov/books/

NBK43682/pd/advances-patterson_48.pd, accessed on January 27, 2012.

26 Agency or Healthcare Research and Quality. (2011) Improving Patient Saety

Through Simulation Research: Funded Projects (AHRQ Pub. No. 11-P012-

EF). Rockville, MD: Retrieved rom http://www.ahrq.gov/qual/simulproj11.

pd, accessed on January 27, 2012. See also Agency or Healthcare Research

and Quality Program Announcement PAR-11-024, Advancing Patient Saety

Through Simulation Research (R18). Retrieved rom http://grants.nih.gov/

grants/guide/pa-fles/PAR-11-024.html, accessed on January 27, 2012.

27 Kimbler W.J. (2008, June 19). Geisinger Medical Center Opens First

Computerized Simulation Training Center in Eastern U.S. Cath Lab Digest.

Retrieved rom http://www.cathlabdigest.com/articles/Geisinger-Medical-


12/13

Center-Opens-First-Computerized-Simulation-Training-Center-Eastern-

US, accessed on January 27, 2012. Preston, P. (2008). Kaiser Permanente:

A Journey in In-Situ Medical Simulation. Presentation rom the Center

or Immersive and Simulation-based Learning, Stanord University.

2008. Retrieved rom http://www.powershow.com/view/a2355-OTNlZ/

Kaiser_Permanente_A_Journey_in_InSitu_Medical_Simulation_ash_ppt_

presentation, accessed on January 27, 2012. Also, or the VA, see United States

Department o Veterans Aairs (2012). SimLEARN Home. Retrieved rom

http://www.simlearn.va.gov/index.asp, accessed on January 27, 2012.

28 Summer, L. Using Health Inormation Technology to Improve Health and

Health Care in Underserved Communities: The Primary Care InormationProject, Health IT or Actionable Knowledge report, AcademyHealth, February

2012.

29 The exception is the NYC PCIP, which is ocused on public and population

health unctions as opposed to clinical research.

30 It is also possible that EHRs can be used as a tool to quickly identiy patients

who meet the initial inclusion criteria or a study. A CTO would then

approach patients to obtain inormed consent and use their own separate

database to record research data.

31 VA Cooperative Studies Program Deputy Director G.D. Huang, personal

communication, January 3, 2012.

32 A recent evaluation o the National Cancer Institutes CaBIG program ound

the programs attempt to bring basic and clinical inormatic tools together in

a single environment was unrealistic given how much the IT needs o these

two types o investigation diverge, see National Cancer Institute Board o

Scientifc Advisers. (2011). An Assessment o the Impact o the NCI Cancer

Biomedical Inormatics Grid (caBIG). Retrieved rom http://deaino.nci.nih.

gov/advisory/bsa/bsa0311/caBIGfnalReport.pd, accessed on January 27, 2012.

33 For example, see Kitano, H. Computational Systems Biology, Nature, Vol. 420,

November 14, 2002, pp. 206-10 and Benson, D. A. et al. GenBank: Update,

Nucleic Acids Research, Vol. 32, Supplement 1, , 2004, pp. D23-6.

34 Presidents Council on Science and Technology, Executive Ofce o the

President. (2008). Priorities or Personalized Medicine. Washington, DC:

White House Ofce o Science and Technology Policy.

35 Ray, T. (2009, October 21). NIH Awards $25M to Kaiser Permanente, UCSF or

100,000-Member Genome-Wide Analysis Data Repository. Pharmacogenomics

Reporter. Retrieved rom http://www.genomeweb.com/dxpgx/nih-awards-

25m-kaiser-permanente-ucs-100000-member-genome-wide-analysis-data-re,

accessed on January 27, 2012.

36 For more detail see United States Department o Veterans Aairs. (2012).

Million Veterans Program (MVP). Retrieved rom http://www.research.va.gov/

mvp/, accessed on January 27, 2012.

37 In general, research requires Institutional Review Board approval, while QI

activities do not, and research is subject to stricter HIPAA privacy restrictions

than is QI. Baily, M.A. et al. QI and Research: Similarities and Dierences,

The Ethics o Using QI Methods to Improve Health Care Quality and Saety.

Hastings Center Special Report, July-August 2006, pp. S11-S21.

38 See or example, Shekelle, P.G. et al. Advancing the Science o Patient Saety,

Annals o Internal Medicine, Vol. 54, No. 10, May 17, 2011, pp. 693-6 and

Clancy, C.M. and D.M. Berwick. The Science o Saety Improvement, Annalso Internal Medicine, Vol. 54, No. 10, May 17, 2011, pp 699-701.

39 Mandl and Lee, 2009, op. cit.; Weiner and Embi, 2009, op.cit.

40 As indicated in an earlier note, the level o quality necessary or even preerred

in EHR data depends on the research question being addressed. When

studying actual medical pract ice, the researcher may actually want to capture

the noisiness o day-to-day care delivery. In other cases, it may be important

or the researcher to know the true clinical value. For example, in studying

the efcacy o a new hypertension medicine, the researcher may want to

know a patients real blood pressure. When studying the eectiveness o a

hypertension management program on strokes or other health outcomes, the

researcher may preer to capture blood pressure measurement as recorded in

actual clinical encounters.

41 Weiner and Embi, 2009, op.cit.

42 Rein, A. Finding Value in Volume: An Exploration o Data Access and Quality

Challenges, Health IT or Actionable Knowledge report, AcademyHealth,

February 2012.

43 Federal policy developed by multiple agencies concerning the protection o

human subjects in research has been codifed in ederal regulations collectively

reerred to as the Common Rule. Additional inormation may be ound

at U.S. Department o Health & Human Services. Federal Policy or the

Protection o Human Subjects (Common Rule). Retrieved rom http://www.

hhs.gov/ohrp/humansubjects/commonrule/index.html,accessed on January 27,

2012.

44 McGraw, D. and A. Leiter. Legal and Policy Challenges to Secondary Uses

o Inormation rom Electronic Clinical Health Records, A Health IT or

Actionable Knowledge report, AcademyHealth, February 2012.

45 Abernathy et al, 2010, op. cit.


13/13

The Potential Impact of Health IT

Documents

Transcript of The Potential Impact of Health IT